Carboxylic acid salts are useful in various applications. For example, salts of iminodiacetic acid may be phosphonomethylated to form N-(phosphonomethyl) iminodiacetic acid (“PMIDA”), which, in turn, may be oxidized to form N-(phosphonomethyl)glycine (known in the agricultural chemical industry as “glyphosate”). See, e.g., Gentilcore, U.S. Pat. No. 4,775,498 (disclosing a method to phosphonomethylate a salt of iminodiacetic acid); Ebner, et al., PCT/US99/03402 (disclosing a method for oxidizing PMIDA). Salts of nitrilotriacetic acid, for example, are excellent chelating agents, and consequently may be used as detergent builders, water-softening agents, scouring aids, dyeing assistants, paper-coating agents, scale inhibitors, and agents for preventing soap degeneration. And many carboxylic acid salts (e.g., salts of glycine, salts of iminodiacetic acid, etc.) may also be neutralized to their corresponding acids and then used, for example, as chelating agents; in food preparations; and as raw materials for making pharmaceuticals, agricultural chemicals, and pesticides. See, eg., Franz, et al., Glyphosate: A Unique Global Herbicide (ACS Monograph 189, 1997) at pp. 234-41 (disclosing the use of glycine and iminodiacetic acid compounds as raw materials to form N-(phosphonomethyl) glycine).
It has long been known that a carboxylic acid salt may be prepared from a primary alcohol by dehydrogenating the alcohol using a copper-containing or silver-containing catalyst. In 1945, Chitwood first reported forming a carboxylic acid salt (specifically, the potassium salt of glycine) by oxidizing a primary alcohol (specifically, monoethanolamine) in an alkaline environment (specifically, in a mixture containing potassium hydroxide) using a copper-containing catalyst (specifically, copper metal or cupric oxide, which reportedly was reduced to copper metal under the reaction conditions) or a silver-containing catalyst (specifically, silver metal or silver oxide, which reportedly was reduced to silver metal under the reaction conditions). See Chitwood, U.S. Pat. No. 2,384,817. Chitwood, however, reported that copper-containing compounds are disadvantageous for this reaction because the copper coagulates over time, thereby causing the copper-containing compounds to have a short duration of maximum catalytic activity. Chitwood also reported that silver-containing compounds have relatively low activity (the silver oxide also reportedly coagulates over time).
In 1988, Goto et al. reported forming a carboxylic acid salt by oxidizing an ethanolamine compound in an alkaline solution (specifically, an aqueous solution containing the hydroxide of an alkali metal or an alkaline earth metal) using Raney copper. See Goto et al., U.S. Pat. No. 4,782,183. Goto et al. reported selectivities of at least 94.8% when dehydrogenating monoethanolamine, diethanolamine, and triethanolamine to form salts of glycine, iminodiacetic acid, and nitrilotriacetic acid, respectively. Raney copper, however, is disadvantageous because (like Chitwood's copper-containing compounds) Raney copper deactivates over time. See, e.g., Franczyk, U.S. Pat. No. 5,292,936, Table 1 (showing the reaction time for Raney copper to increase from 4 to 8 hours over 9 cycles).
Various developments have been reported which address the instability of copper-containing catalysts when used to dehydrogenate primary alcohols. Although these developments have made the use of copper catalysts more commercially viable, their results are still not entirely satisfactory.
Franczyk, for example, reports that copper-containing catalysts particularly Raney copper) can be stabilized by using such a catalyst which also contains 50 to 10,000 parts per million of one or more various other metals selected from the group consisting of chromium, titanium, niobium, tantalum, zirconium, vanadium, molybdenum, tungsten, cobalt, nickel, bismuth, tin, antimony, lead, and germanium, with vanadium, chromium, and molybdenum being the more preferred metals. See Franczyk, U.S. Pat. Nos. 5,292,936; 5,367,112; and 5,739,390. Although such metals do tend to impart a stabilizing effect to a copper catalyst, this effect often decreases over time. See, e.g., Franczyk patents, Table 2 (showing the reaction time decreasing from 5.8 hours to 8.0 hours over 25 cycles) and Table 4 (showing the reaction time decreasing 3.1 to 5.5 hours over 12 cycles). This decrease is due, at least in part, to the fact that such metals tend to leach over time as the catalyst is used, particularly where the primary alcohol or the dehydrogenation product is a chelating agent (e.g., a salt of iminodiacetic acid).
Ebner et al. report using a catalyst comprising copper supported on an alkali-resistant support (particularly a carbon support) to dehydrogenate primary alcohols to make carboxylic acid salts. See Ebner et al., U.S. Pat. No. 5,627,125. This catalyst also comprises about 0.05 to about 10% by weight of a noble metal to anchor and disperse the copper to the support. Although Ebner et al. report shorter reaction times with their catalyst relative to previously disclosed copper-containing catalysts, their catalyst is costly due to the need for the noble metal to anchor the copper to the support. In addition, the added volume of Ebner et al.'s catalyst due to the carbon support can, in some instances, make handling the catalyst cumbersome, consequently reducing throughput. Further, Ebner et al.'s catalyst often loses activity over time with use (although the rate of deactivation is often less than the rate of deactivation of the Franczyk catalysts). See, e.g., Ebner et al., Table 1 (showing the reaction time increasing from 103 to 150 minutes over 9 cycles) and Table 2 (showing the reaction time increasing from 61 to 155 minutes over 8 cycles). As with the Franczyk catalysts, this problem tends to arise particularly where the primary alcohol or the dehydrogenation salt product is a chelating agent.
Other reported copper-containing catalysts contain a non-carbon support, such as, SiO2, Al2O3, TiO2, ZrO2, and the like. See, e.g., Akzo Nobel, WO 98/13140 (disclosing a catalyst consisting of copper on ZrO2). These supports, however, tend to be vulnerable to attrition under the reaction conditions normally present when dehydrogenating a primary alcohol, and are therefore usually less suitable than Ebner et al.'s carbon supports. This vulnerability to attrition tends to also cause these supports to exhibit poor filtration characteristics.
Use of copper-containing and silver-containing catalysts in other types of oxidation reactions has also been reported. Applicants, however, are unaware of any such disclosures which address the problems associated with copper-containing or silver-containing catalysts in processes involving the dehydrogenation of primary alcohols to form carboxylic acid salts.
Boumonville et al. report forming a ketone by dehydrogenating a secondary alcohol using a Raney nickel catalyst containing 0.1 to 10% by weight of copper, silver, gold, tin, lead, zinc, cadmium, indium, or germanium. See Bournonville et al., U.S. Pat. No. 4,380,673. This reaction, however, does not form a carboxylic acid salt—forming a carboxylic acid salt would further require the cleavage of an alkyl group from the carbonyl group and the subsequent attachment of a hydroxy salt to the carbonyl group. In addition, Bournonville et al. report that their reaction is catalyzed by the Raney nickel, and that the function of the additional metal (e.g., copper or silver) is to suppress hydrogenolysis side reactions. See Boumonville et al., col. 3, lines 45-47. This is in contrast to dehydrogenation reactions of primary alcohols using copper catalysts, such as Raney copper, where catalytic activity is provided primarily by copper atoms near the surface of the catalyst.
Yamachika et al. report forming benzaldehydes by reducing benzonitriles in the presence of acid and a Raney nickel catalyst which has been pre-treated with a copper salt solution. See Yamachika et al., U.S. Pat. No. 4,500,721. Yamachika et al. disclose that the conditions of catalyst pre-treatment should be sufficient to form a catalyst which contains 5 to 80% (more preferably 10 to 60%) by weight of copper. Yamachika et al. report that the presence of the copper increases the yield of benzaldehydes during the reaction. This reaction, however, is conducted in an acidic environment, is not directed to dehydrogenating primary alcohols (or any other alcohols), and does not form carboxylic acid salts.
Thus, although positive advances have been reported for converting a primary alcohol to a carboxylic acid salt using a copper-containing catalyst, there continues to be a need for a more economical liquid-phase process which uses a catalyst that has high surface area, has high activity, and exhibits stability (i.e., maintains its activity) over time with usage. This need particularly exists where the primary alcohol substrate and/or carboxylic acid salt product is a chelating agent (e.g. salt of iminodiacetic acid).
The hydrogen produced by the dehydrogenation of primary alcohols can also be useful, particularly in the production of fuel cells. For example, W. H. Cheng, in Acc. Chem. Rev., vol. 32, 685-91(1999), describes the conversion of primary alcohols such as methanol to hydrogen as a safe and readily transportable source of hydrogen fuel cells for a variety of applications, most notably automotive applications. Thus, the more economical liquid-phase process of the present invention for the dehydrogenation of primary alcohols can also lead to more economical production of hydrogen from primary alcohols.